I believe that the more complex / convoluted the nipping / gripping nub becomes, the less efficiently it can nip / grip a penetrating line. In a complex / convoluted nub, a considerable portion of its overall constricting power will be "wasted' within itself. Some segments of the nub which surround the penetrating line will not be directly pulled and tensioned by the Standing End, and so they will not squeeze the penetrating line very much. On the contrary, some of them will form a kind of a "protective shield" around it, and they will absorb a part of compression which otherwise ( in a simpler, less tangled but more tightly closed around itself nub ) would had been able to reach it directly, and press/squeeze them a more unobstructed, so more efficient way.
Multiplying the points and increasing the total area of contact between the nipping / gripping nub and the penetrating line, is a poor strategy. In fact, what we should better do, is the exact opposite : fewer points, smaller area, so greater, more efficient concentration of the pressure applied by the nipping / gripping nub on the surface of the penetrating line. This way the cylindrical surfaces of the adjacent segments will be squeezed locally, they will settle into a saddle-shaped forms , and this will multiply the friction forces between them. Otherwise, the two segments will run the danger to slide on each others flat surface, and the total friction forces will be reduced.
The great example of this is the nipping / gripping efficiency of the opposed bights locking mechanism : Few points and small area of contact - but of unobstructed, direct contact, between a short segment of the penetrating line and the fully tensioned segments at the first curves / tips of the two bights.
Of course, the shape of the path of the penetrating line is of paramount importance, too. An L-shaped path, plays the role of a "handle", by which the nipping / gripping nub can 'hold" the penetrating line more easily. ( A U-shaped path is even more effective, but it can not be used in adjustable loops, because either it will prevent any pull of the Tail End during the adjustment of the size of the eye, or it will force the nub itself to be deformed badly during this adjustment, and allow the U to become straightened out, and release the penetrating line. )
Another way we can increase friction on the penetrating line, is, simply, to use a knot that we know it jams ! I have seen that the humble Clove hitch, when tied around a segment of a compressible material / a rope, "closes" around itself very tightly, and "works" as a tensile forces accumulator, a rope-made ratchet which absorbs all induced through its ends tensile forces, but then "locks" them securely inside its two wraps.
In short, to increase friction, we should either increase quantity of friction= increase the number of helical turns around a nipping / gripping coil, just as we do in climbing hitches - OR increase quality of friction = reduce the complexity of the nipping / gripping nub, reduce the number of points and the area of contact between this nub and the penetrating line, but :
1. Be sure that the segments of the nub which come in contact with the penetrating line are at the direct continuation f the Standing End, and so maximally loaded / tensioned. It is the first curve which does most of the job, the segments of the nub after this play a secondary, only, role.
2. The penetrating line follows an L-shaped path within the surrounding nub.
